GridARM: an automatic runtime management framework for SAMR applications in grid environments

Abstract

Structured adaptive mesh refinement (SAMR) techniques offer the potential for accurate solutions of physically realistic models of complex physical phenomena, and are being effectively used in many domains including computational fluid dynamics, numerical relativity, astrophysics, subsurface modeling and oil reservoir simulation. However, the inherent space-time heterogeneity and dynamism of SAMR applications coupled with a similarly heterogeneous and dynamic execution environment such as the computational Grid present significant challenges in application composition, runtime management, optimization, and adaptation. This paper presents the design of GridARM – an autonomic runtime management framework that monitors application and system state and provides appropriate distribution, configuration, scheduling, and adaptation strategies to optimize performance and support the efficient and scalable execution of SAMR implementations in Grid environments.